Motivation

This project was created with the intention of becoming the web scraping library for node. My process for the development of this project begins by brainstorming a definitive answer for the way things should work, rather than simply writing code to make things work.

Experience

Being my first open source project, I learned a lot while developing Osmosis. Very early on I learned how absolutely necessary unit testing is for maintaining the stability of any code base. To better ensure that no regressions are made in the future, I've written comprehensive unit tests for every feature in Osmosis.

Osmosis also helped me learn more about Git and the Git work flow, particularly the importance of having a sane Git model for every repository. During the early development of Osmosis, I switched to git-flow, which streamlined all aspects of the development cycle and helps to keep Osmosis in a sane, maintainable environment.

Motivation

This project is used by Osmosis for fast and lightweight XML/HTML document parsing. Because Osmosis depends on it, I've come to be a major contributor maintaining, developing, and debugging this library.

Experience

In addition to learning a little bit more about how to maintain the quality and stability of open source code, contributing to this project has resulted in a deeper knowledge of NodeJS and V8 internals. Writing a native module in C++ has given me more insight into things like memory management and garbage collection.

For example, my first few contributions fixed many long-standing memory leak problems. The memory leaks were very noticeable and the existing memory management solution wasn't sufficient.

Additionally, I've contributed many new libxmljs API feature, helped to manage and fix any issues that have arisen, review and merge pull requests, and maintain portability across platforms, including the support of pre-built binaries for Windows users.

These days I'm the primary maintainer of the project. I've been working on a complete rewrite of the library, which includes features like:

• Backwards compatible API abstractions rewritten in Typescript
• Auto-generated Typescript definitions
• Auto-generated bindings (using SWIG)
• Auto-generated documentation (using jsdoc)
• Support for multi-threaded parsing

Motivation

This project was created to enable a browser-like interaction with web pages using libxmljs objects. It's used by Osmosis for document parsing and searching as well as the ability to run embedded javascript code found on web pages.

The benefit of this project over other javascript DOM implementations is that it's based on native bindings and wrappers around native libxml objects. This makes document parsing, searching, and manipulation significantly faster than pure JS implementations. It allows interacting with enormous numbers of web documents with native speed, low overhead, and minimal resource consumption.

Experience

I've been told that creating a DOM implementation seems a bit ambitious. While that may be true, I knew the benefits of wrapping libxmljs, so I decided to go for it anyways.

Writing a DOM wrapper implementation for libxmljs has not only benefitted Osmosis, but has also given me a much deeper knowledge of the DOM API standards.

Motivation

This project was created for Osmosis and libxmljs-dom to give them the ability to use CSS selectors with Libxmljs. Libxmljs uses libxml, which only uses XPath selectors for node selection.

The idea is not just to allow conversion of standard CSS selectors to valid XPath selectors. There are two important additional features that this library implements.

First, css2xpath implements common CSS selector extensions, such as those used by jQuery. Second, css2xpath allows combining CSS and XPath selectors in the same expression, giving you all the power of XPath selectors, with the simplicity and elegance of CSS selectors.

Experience

While creating this project I've learned a lot more about the CSS and XPath standards. This project has given me the opportunity to take a deep dive into CSS and XPath selector syntax and functionality.

Motivation

This project was created so that I could quickly prototype and test machine learning models in javascript, even on large datasets.

The existing libraries for hierarchical clustering were either inefficient or had APIs, options, and outputs that demonstrated a lack of understanding on the concept.

This project was written using optimal javascript for this use-case and has several algorithmic optimizations over other implementations.

For example, agglo uses a linked-list to iterate and update the mutable "linkage" matrix, whereas other algorithms will traverse the entire array or use a computationally expensive array-slice.

This project also provides an API that makes it simple to implement an out-of-the-box clustering model, while still allowing full customization and integration into the clustering process.

Experience

While creating this project I've learned a lot about optimizing javascript code so that it runs well in different execution environments.

Motivation

This project was created to provide a traditional C-style approach to building out comprehensive machine learning models using native code.

While there are other C/C++ machine learning libraries that provide agglomerative clustering functionalities, these implementations have several downsides.

For example, these implementations require the use of the parent library's formats and data structures, and are often unoptimized, while lacking flexibility or even understanding of the algorithm itself.

The project's goal was to create a more format/structure agnostic approach that could fit nicely into to any existing C/C++ codebase with a minimum knowledge, while still being really fast and memory efficient.

This implementation also has several algorithmic optimizations over existing algorithms.

For example, libhcluster uses a linked-list to iterate and update the mutable "linkage" matrix, whereas other algorithms will traverse the entire array or use a computationally expensive array-slice.

The implementation is also designed to use the smallest memory footprint possible for a full-fledged hierarchical agglomerative clustering API with complete customizability.

Experience

While creating this project I've become more familiar with memory management, callbacks in C, and C-style API standards.

Motivation

The motivation for this project was to simplify common Wordpress integrations into single "data-" attributes, allowing generic HTML/CSS UI components to be transformed into a fully functional Wordpress theme.

Currently, the primary use of this project is to enable a standardized way to build Wordpress themes using professional WYSIWYG tools like Webflow or Pinegrow.

Experience

While creating this project I've grown to further understand the difficulty in maintain backward compatibility and appropriate versioning between an API and command line tool.

Motivation

This project initially started as an experimental proof-of-concept, prompted by the question "if the server and client both speak javascript, can't we write our app as single code base and have the execution switch endpoints as-needed?"

Motivation

The motivation for this project was to create a javascript runtime that could be used with Symmetric, another project of mine.

In order to be compatible with Symmetric, there were two primary requirements:

1. to be able to pause and resume execution at any point
2. to run in the browser (i.e. written in high-compatibility javascript)

While there are many other examples of javascript evaluators written in javascript that consume an ESTree-compatible AST, none if them suited the needs required.

For example, most of them are outdated or missing implementation of vital javascript language functionality.

Experience

Since this project requires rewriting every javascript langauge feature (in javascript), I've learned a lot about javascript as a language.